PSI - Issue 84

Vincenzo Gattulli et al. / Procedia Structural Integrity 84 (2026) 41–48

45

4. Digital Twin Framework and Data Flow The DT implementation for LNGS Hall B integrates a high-fidelity geometric model with a distributed sensor based monitoring system. These components are complementary: the geometric model provides the spatial reference for contextualizing measurements, while the monitoring system delivers continuous data on structural and environmental conditions. Similar approaches have been applied to heritage structures, where point cloud–based HBIM models are used to generate analysis-ready finite element models and improve the representation of structural behavior through data-informed assumptions (Talebi et al., 2026). 4.1. Geometric Data Acquisition, Georeferencing, and Spatial Consistency The geometric characterization of Hall B is derived from multiple three-dimensional survey techniques adopted to address the constraints of the underground environment. Static terrestrial LiDAR scanning provides high-accuracy geometry in accessible areas, while mobile SLAM-based mapping enables efficient coverage of elongated sections where conventional surveying is less practical. In addition, oblique photogrammetry is used to capture localized details in areas affected by visibility or accessibility limitations (Menna et al., 2014). The integration of these complementary techniques enables the generation of dense point clouds at varying levels of detail, which are aligned to obtain a coherent geometric representation suitable for Digital Twin applications. Due to the absence of GNSS signals underground, the datasets are referenced using survey control derived from total station measurements. GNSS-tied total station tracks provide a consistent spatial framework for aligning LiDAR, SLAM, and photogrammetric point clouds within a common reference system (Bayer et al., 2018). This georeferencing strategy ensures spatial consistency across datasets acquired at different times and with different techniques, supporting long-term monitoring and accurate spatial localization of sensor data within the Digital Twin environment. 4.2. Monitoring Infrastructure and Measured Quantities The monitoring infrastructure consists of a distributed network of IoT-enabled sensors deployed along Hall B to capture structural and environmental parameters relevant to performance and operational safety. The system integrates different sensor types, including accelerometers for dynamic response, inclinometers and extensometers for deformation monitoring, and environmental sensors measuring temperature, humidity, and air-related parameters. Sensor deployment ensures continuous operation under the strict constraints of the LNGS facility, including limited accessibility and sensitivity to environmental disturbances. Data are acquired locally, transmitted to the gateway, and then to the centralized platform described in Section 3 for storage and analysis. An overview of the monitored quantities and corresponding sensor categories is reported in Table 1.

Table 1. Overview of monitored quantities and sensor categories adopted for the Digital Twin system at LNGS Hall B

Sensor Category

Measured Quantities

Purpose within the Digital Twin

Structural monitoring

Acceleration time histories

Characterization of dynamic response and vibration behaviour Assessment of environmental stability and operating conditions

Environmental monitoring

Temperature, humidity, and air-related parameters Survey control points and reference tracks

Geometric reference

Spatial consistency and alignment of geometric and monitoring data

4.3. Integration of Geometry and Monitoring Data The high-fidelity geometric model serves as the spatial backbone of the Digital Twin, allowing each sensor measurement to be associated with a precise location within the underground structure and supporting coherent visualization and interpretation of monitoring results. By integrating geometric representation with continuous sensor-based monitoring, the platform enables spatial mapping of measured quantities onto the structural domain, facilitating the identification of localized phenomena